Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 8 Jul 2010 14:52:15 +0530
From:      "Jayachandran C." <c.jayachandran@gmail.com>
To:        "M. Warner Losh" <imp@bsdimp.com>
Cc:        freebsd-mips@freebsd.org
Subject:   Re: Merging 64 bit changes to -HEAD - part 4
Message-ID:  <AANLkTilFZ4zAKqGcDKie6og5Nhrh9PU_Ta3uDoq7yexA@mail.gmail.com>
In-Reply-To: <20100708.021250.1099368555950605809.imp@bsdimp.com>
References:  <AANLkTikSVi27V2UICgLvKd8Bk7v6tuGty9YX6-C6-21H@mail.gmail.com> <20100708.021250.1099368555950605809.imp@bsdimp.com>

next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, Jul 8, 2010 at 1:42 PM, M. Warner Losh <imp@bsdimp.com> wrote:
> In message: <AANLkTikSVi27V2UICgLvKd8Bk7v6tuGty9YX6-C6-21H@mail.gmail.com=
>
> =A0 =A0 =A0 =A0 =A0 =A0"Jayachandran C." <c.jayachandran@gmail.com> write=
s:
> : On Tue, Jun 15, 2010 at 7:06 PM, Jayachandran C.
> : <c.jayachandran@gmail.com> wrote:
> : > I have volunteered to merge Juli's 64-bit work into HEAD, =A0and
> : > hopefully get it to work on XLR too. The tree
> : > (http://svn.freebsd.org/base/user/jmallett/octeon) has quite a bit of
> : > changes, so I would like to do this over multiple changesets and
> : > without breaking the current o32 code.
> :
> : Here's the next installment, this has the next set of Juli's changes
> : and some fixes to get it working on XLR. The patches in this set are:
> :
> : mips-segtab-macro.patch :
> : =A0 Change PDE/PTE access to a macro (from Juli's branch)
>
> This looks ok, I think.
>
> : mips-cache-fix.patch :
> : =A0Minor fix for cache code (JC)
>
> This looks good.
>
> : rmi-other.patch:
> : =A064 bit compilation fixes for sys/mips/rmi (JC)
> : =A0Fixes to platform and driver code for 64 bit compilation, including
> : changes to ethernet driver.
>
> This is xlr specific, so I can't comment.
>
> : mips-rmi-kx-enable.patch:
> : =A0Changes to enable KX bit for TARGET_XLR_XLS (JC).
> : =A0I have added another case for the TARGET_OCTEON #ifdef in exception.=
S
> : and locore.S, but I think this can be moved to =A0a header file later.
>
> Yea, we need to do some rework here, but I think this is basically OK.
>
> : pmap-n64.patch
> : =A0The main n64 patch (from Juli's branch)
> : =A0 This still uses the old 2-level page tables. But this adds other
> : pmap code to support n64. I have re-arranged some of =A0Juli's code to
> : reduce #ifdefs.
>
> I think this could be done in smaller bites. =A0At least the
> MIPS_{PYHS,KSEGx}_TO_{KSEGx,PHYS} stuff can be done as a separate
> patch.
>
> I really like the use of direct maps where possible for 64-bit
> machines.
>
> There's a number of comment fixes in this, which seem to be logically
> independent as well. =A0Not sure it is worth the effort of busting these
> out though.
>
> : runq-64.patch
> : =A0 64-bit rqb_word_t for n64 (JC)
>
> I think this is cool.
>
> : ldscript-64.patch
> : =A064 bit linker script (JC)
> : =A0linker script for 64 bit compilation, and XLR configuration file.
>
> I wonder if the octeon ld script can be merged or if we can use this
> for octeon. =A0Some investigation is warranted, I think...
>
> : With these changes, a n64 kernel can be compiled and it will to
> : 'mountroot>' on XLR. =A0 There is more code left to merge in Juli's
> : branch - 32 bit compat code, sfbuf/uio, UMA alloc, and DDB. which I
> : should be able to get in the next =A0one or two passes.
>
> OK.
>
> : Let me know your comments. There are only two minor changes should
> : affect the existing o32 code paths in the above code (moving the check
> : for >512M, and change in pmap_map to handle KSEG0 addresses), but let
> : me know if something breaks.
>
> Will do. =A0I'll see if I can create a system with these patches...

Thanks for the review, I'll try to get the patches other than pmap-n64
in first (with mentor approval). I'll post a new split version of
pmap-n64 after that.

JC.



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?AANLkTilFZ4zAKqGcDKie6og5Nhrh9PU_Ta3uDoq7yexA>